Using a standardized AIDS prevention outreach program as a baseline, we propose a field experiment to test the effectiveness of a "user-driven" approach to distributing AIDS prevention materials and recruiting injection drug users (IDUs) for intervention. The two study areas each contain between 300 - 500 IDUs (a more scientifically-based measurement is one goal of the research), and the field experiment will assess the ability of each program to reach and change the behavior of IDUs in both areas. Traditional outreach relies on a professional staff, drawn primarily from recovered users, to recruit current users into AIDS prevention programs, and to distribute AIDS prevention materials and information. The user-driven intervention proposed here is based on the theory of group-mediated social control. The program will provide IDUs with incentives to recruit other IDUs into a standardized, multicultural interview-test-education session, to distribute prevention materials such as bleach and condoms among peers, and to disseminate AIDS prevention information. It will be based on an expanding system of chain-referrals that is robust enough to saturate the IDU population. The aim is to coopt user networks to serve as a medium for transmission of risk-- reduction materials and information, and to recruit IDUs for education about AIDS, HIV, treatment options and alternatives to their present lifestyle. This contrasts with the traditional approach which relies on a staff of paid outreach workers to generate referrals and distribute prevention information and materials. A user-driven intervention has three potential advantages over traditional out-reach. First, since the former relies on current user networks for recruitment, its coverage of the target population may be more extensive than traditional outreach projects. Second, given that user-driven outreach will actively cultivate and rely upon the persuasion and influence from peers rather than from a professional staff, it may be more effective at securing user compliance. Third, a user-driven intervention may avoid many of the problems, risks and costs which outreach projects face in hiring recovered users to work in high crime, heavy drug using scenes. Owing to reduced reliance on a professional staff, the cost of a user-driven intervention may be only a seventh of that incurred by traditional outreach programs.